We need to save off the full 64-bit register, not just the low 32 bits,
of all registers getting saved off in _rtld_bind_start. Additionally,
we need to save off the other SPE registers (SPEFSCR and accumulator),
so that their program state is not affected by the PLT resolver.
Details
Details
Diff Detail
Diff Detail
- Repository
- rS FreeBSD src repository - subversion
- Lint
Lint Not Applicable - Unit
Tests Not Applicable
Event Timeline
| libexec/rtld-elf/powerpc/rtld_start.S | ||
|---|---|---|
| 114 ↗ | (On Diff #64766) | Comment is outdated now. Maybe something like /* stack space for r0, r3-r31 and nonvolatile SPRs + callee save area */ |
| 126 ↗ | (On Diff #64766) | same here. Probably don't need to duplicate the comment either since it's now obvious what it's doing. |
| libexec/rtld-elf/powerpc/rtld_start.S | ||
|---|---|---|
| 110 ↗ | (On Diff #64766) | We're now allocating 20 bytes less stack than the original code in the !SPE case. Your math looks more correct to me though, so the original -160 must have been wrong. |
| libexec/rtld-elf/powerpc/rtld_start.S | ||
|---|---|---|
| 110 ↗ | (On Diff #64766) | Yes, it was overkill, but not detrimental. |
Comment Actions
Attempts to theorize on reduction of scope of this have failed, apparently we do in fact need to be backing this much up when calling the resolver.
I think this is about as nice as we're gonna get it.